In its journey from blood vessels to cancer cells a therapeutic agent must overcome the transport barrier posed by the interstitial matrix. In contrast to our understanding of blood vessels, our understanding of the formation and function of the tumor interstitial matrix is still in its infancy. To make progress toward the goal of normalizing interstitial transport, we need to (i) identify the matrix components that contribute to the barrier function of the interstitium, (ii) image these components in vivo, (iii) measure interstitial transport parameters with high sensitivity, and (iv) modify the matrix without harming the host. In the past 5 years we have made significant progress toward these goals. We discovered that collagen is the key determinant of the interstitial barrier (Cancer Research 2000), established the use of second harmonic generation for in vivo imaging of the collagen matrix in tumors (Nature Medicine 2003), adapted multi-photon fluorescence correlation spectroscopy in vivo and revealed the two phase nature of interstitial transport in tumors (Nature Medicine 2004a), and used an endogenous molecule, the hormone relaxin, to modify the collagen matrix and make it more penetrable (Nature Medicine 2003). Finally, we demonstrated that the xenograft models can be used to estimate diffusion in human tumors. This project builds upon these exciting technical developments and scientific findings. Our new goal is to reveal the molecular underpinnings of matrix modification and to identify additional, novel matrix modifiers that can be used clinically. TGF-fJl and PDGF can stimulate production of collagen by stromal cells in tumors;therefore, we will determine if modulating these molecules or their downstream signaling pathways improves interstitial transport (Aim1). Since MMP-1, -8, and -13 degrade fibrillar collagen, we will express these.MMPs in cancer cells or stromal cells, or deliver them with a replication-defective virus, and measure transport (Aim 2). Finally, we will determine the mechanism and extent to which relaxin and halofuginone - two agents proven to be safe in humans - increase the delivery and efficacy of clinically approved therapeutics (Aim 3). We will carry out these studies in orthotopically-growing human mammary carcinoma and melanoma xenografts as well as in a spontaneous tumor model using IgG, IgM, AAV, HSV-1 and Herceptin as models of macromolecular therapeutic agents. We have the necessary technology, molecular reagents and animal models from Cores A, B and C (Nature Reviews Cancer 2002) as well as from our collaborators. We also have clinicalcollaborators in place to initiate a clinical trial based on our findings, similar to our ongoing trial using a VEGF-specific antibody (Nature Medicine 2004b).